U.S. patent number 3,602,703 [Application Number 04/867,541] was granted by the patent office on 1971-08-31 for power demand predicting control system.
This patent grant is currently assigned to International Business Machines Corporation. Invention is credited to Thomas R. Polenz.
United States Patent |
3,602,703 |
Polenz |
August 31, 1971 |
POWER DEMAND PREDICTING CONTROL SYSTEM
Abstract
A power consumption control system and method for controlling
total power consumption during a demand billing period of a
plurality of variable rate consuming devices which operate in a
known pattern. The devices are monitored to determine the stage of
operation of the known pattern for each device. The devices are
allocated priorities according to predetermined classifications.
The total power consumption of the highest priority devices is
estimated, based upon the power consumption of the present and
immediately following stages of operation thereof and the expected
time to be spent in each, for the remainder of the demand billing
period, and that amount of power is allocated thereto. The maximum
power consumption of the remaining devices over the same period is
likewise estimated and the remaining available power is allocated
thereto in accordance with the ratio of the available power to the
maximum power consumption thereof.
Inventors: |
Polenz; Thomas R. (Greendale,
WI) |
Assignee: |
International Business Machines
Corporation (Armonk, NY)
|
Family
ID: |
25349981 |
Appl.
No.: |
04/867,541 |
Filed: |
October 20, 1969 |
Current U.S.
Class: |
700/291 |
Current CPC
Class: |
H02J
3/14 (20130101); H02J 2310/64 (20200101); Y02B
70/3225 (20130101); Y04S 20/222 (20130101); Y04S
50/10 (20130101) |
Current International
Class: |
H02J
3/12 (20060101); H02J 3/14 (20060101); G05b
013/02 (); G05b 015/02 (); G06f 015/46 () |
Field of
Search: |
;235/151.21 |
Other References
Tippett, J. et al. Controlling Megawatts in Steelmaking, In Control
Engineering, p. 68-70, June, 1965, TJ212.C6..
|
Primary Examiner: Botz; Eugene G.
Assistant Examiner: Dildine, Jr.; R. Stephen
Claims
What is claimed is:
1. Apparatus for controlling total power consumption during a
demand billing period of a plurality of controllable variable-rate
power-consuming devices which operate in a known pattern, said
devices including control means operable to separately control the
power consumption rate of each said device and thereby stretch or
shorten said pattern therefor, and wherein metering means supplies
signals indicative of the power consumption of said devices
comprising:
available power computer means responsive to said signals from said
metering means and to a preset limit, which limit indicates the
desired limit of power consumption during said demand billing
period, for periodically computing the amount of remaining power
available for consumption by said devices during the remainder of
said demand billing period;
priority means for periodically allocating priorities between said
devices in accordance with a predetermined classification;
first power usage computer means responsive to said priority means
for determining the point in said pattern of said devices having
the highest priority allocated thereto and responsive to said
determination to estimate the power consumption thereof as defined
by said pattern for the remainder of the demand billing period, and
for signalling said control means therefor to operate said
associated device at the maximum available power consumption rate
in accordance with said pattern,
subtraction means responsive to said first power usage computer
means and said available power computer means for subtracting the
value of said estimated power consumption from the value of said
remaining available power to thereby determine the amount of power
available for the remaining ones of said devices; and
Second power usage computer means responsive to said priority means
for determining the point in said pattern of each of said devices
not having the highest priority allocated thereto and responsive to
said determination to estimate the maximum amount of power
consumption thereof as defined by said pattern for the remainder of
the demand billing period, additionally responsive to said amount
of power available for said remaining devices and to said estimated
maximum amount of power consumption thereby to establish a ratio
therebetween, and for signalling said control means therefor to
operate said associated device at a power consumption rate in
accordance with said ratio as related to the maximum rate of said
pattern.
2. The apparatus in accordance with claim 1 wherein:
said first power usage computer means includes means responsive to
the occurrence of said signalling of said control means thereby for
updating said point determination therein for each said remaining
device in accordance with said pattern at the power consumption
rate in accordance with said ratio.
3. The apparatus in accordance with claim 2 wherein:
said first and second power usage computer means each include means
to estimate said power consumption of each said associated device
by employing the updated point in said pattern for estimating the
power consumption rates and time to be spent at each such rate for
each said device during the remainder of said demand billing period
and employing said estimated rates and times to estimate the
maximum total power consumption of said associated devices during
the remainder of the demand billing period.
4. The apparatus in accordance with claim 3 wherein:
said power consumption estimation means in said power usage
computer means includes multiplication means for multiplying the
value of each said power consumption rate for each associated
device by the value of said time to be spent thereby at such rate
during the remainder of said demand billing period and adding means
to add the products of said multiplication.
5. The apparatus in accordance with claim 4 additionally
including:
timing means responsive to a synchronizing signal supplied from a
synchronizing source to thereupon and at predetermined periodic
intervals supply signals to said available power computer means and
said priority means to cause said periodic operation thereof.
6. The apparatus in accordance with claim 5 additionally
including:
monitoring means for monitoring the power consumption of each of
said devices to determine the point in said pattern of each said
device and responsive to said timing means for periodically
supplying signals to said first and second power usage computer
means for correcting said point-determining means. 7The method of
controlling total power consumption during a demand billing period
of a plurality of controllable variable-rate power-consuming
devices which operate in a known pattern, said devices including
control means operable to separately control the power consumption
rate of each said device and thereby stretch or shorten said
pattern therefor, and wherein metering means supplies signals
indicative of the power consumption of said devices, comprising the
following series of steps, which series is initiated at
predetermined intervals;
responding to said signals from said metering means and to a preset
limit of the amount of power consumption for said demand billing
period to compute by an instrumentality the amount of remaining
power available for consumption by said devices during the
remainder of said demand billing period;
allocating priorities between said devices in accordance with a
predetermined classification;
based upon an estimate of the present point in said pattern of the
operation of said devices having the highest priority allocated
thereto, computing by an instrumentality an estimate of the power
consumption thereof as defined by said pattern for the remainder of
the the demand billing period;
signalling by an instrumentality said control means for said
highest priority devices to operate said associated device at the
maximum available power consumption rate in accordance with said
pattern:
subtracting by an instrumentality the value of said estimated power
consumption from the value of said remaining available power to
thereby determine the amount of power available for the remaining
ones of said devices;
based upon an estimate of the present point in said pattern of
operation of said devices not having the highest priority allocated
thereto, computing by an instrumentality an estimate of the maximum
amount of power consumption thereof as defined by said pattern for
the remainder of the demand billing period;
computing by an instrumentality the ratio between the value of said
amount of power available for said remaining devices and the value
of said estimated maximum amount of power consumption thereby
and
signalling by an instrumentality said control means for said
remaining devices to operate said associated device at a power
consumption rate in accordance with said ratio as related to the
maximum rate of said pattern.
. The method in accordance with claim 7 additionally including the
steps of:
updating by an instrumentality said estimate of said point in said
pattern of the operation of each said highest priority device in
accordance with said pattern at said maximum power consumption rate
thereof; and
updating by an instrumentality said estimate of said point in said
pattern of the operation of each said remaining device in
accordance with said pattern at the power consumption ratio thereof
in accordance with said
ratio. 9. The method in accordance with claim 8 wherein said steps
of computing estimates of the power consumption of said devices
each comprise the steps of:
based upon said updated estimate of the present point in said
pattern of the operation of each of said associated devices,
computing by an instrumentality an estimate of the power
consumption rates and time to be spent at each such rate for each
said device during the remainder of said demand billing period;
and
computing by an instrumentality from the values of said estimated
rates and times an estimate of the maximum total power consumption
of said
associated devices during the remainder of the demand billing
period. 10. The method in accordance with claim 9 wherein said step
of computing an estimate of the maximum total power consumption of
said associated devices comprises the steps of:
multiplying by an instrumentality the value of each said estimated
power consumption rate for each associated device by the value of
said estimated time to be spent thereby at such rate; and
adding by an instrumentality the products of said multiplication.
11. The method in accordance with claim 9 including the additional
step of:
monitoring by an instrumentality the power consumption of each of
said devices to determine the point in said pattern of each said
device; and
correcting by an instrumentality said estimates of said point in
pattern of each said device in accordance with said determination
thereof.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to power consumption control and, more
particularly, to prediction and control of power consumption during
a demand billing period.
2. Description of the Prior Art
Electric utility companies normally employ a demand billing
technique for the billing of large customers. Under this system,
electric utilities typically divide the day into 15 minute demand
billing time periods. The billing rates increase substantially if
the customer consumes electric power irregularly throughout the
day. Hence, it is to the customer's advantage not to use more power
during one 15 -minute time period than he uses during any other
such time period.
Attempts have been made in the past to control electrical power
according to a demand billing scheme. Typically, these techniques
examine power usage during an immediately preceding short period of
time, such as 1 or 2 minutes, and assume that this rate of power
use will continue during the remainder of the demand period. Such
systems project these values through the remainder of the demand
period and control the load in accordance with this projected or
anticipated demand. In this manner, these systems calculate the
total amount of power that would be used during the remainder of
the demand period if the current use continues, and compare this
amount to the amount of power available according to the demand
limit. If this calculation shows that the limit will be exceeded,
some control action is taken to reduce the rate of power
consumption. Hence, these systems are concerned primarily with
maintaining the load less than the predetermined demand limit and
thus do not utilize the maximum available load throughout the
demand period. Accordingly, power which might be consumed at a
cheaper rate is not consumed since the load anticipation does not
contain provision for maintaining the highest possible load factor
on the system.
Various improvements have been made to this basic method to some
what increase the accuracy thereof. For example, control systems
have been arranged to compute the average rate at which power is
consumed over progressively smaller subperiods during the demand
period. The predicted demand at the end of each subperiod is
derived by adding accumulated power already actually consumed to
the product of the present computed average rate times the time
remaining in the demand period. Thus, a brief duration high or low
rate of power consumption may be allowed to average out during the
first part of the demand period due to the relatively long
subperiods. Still, however, cutbacks in the rate of power
consumption made because of power consumption extrapolation may not
be recouped later in the demand period when various ones of the
power-consuming devices progress to new stages of operation having
reduced or no power consumption. Again, efficiency of operation
through full utilization of the equipment is not accomplished and
power which might be consumed at a cheaper rate is not
consumed.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to allow
prediction and control of power consumption of variable-rate
consuming devices which maintains the highest possible load factor
on the system consistent with a constant consumption on a periodic
demand billing basis.
Briefly, the invention comprises a system and method for
controlling total power consumption during a demand billing period
of a plurality of variable-rate consuming devices which operate in
a known pattern. Monitoring means is provided to monitor each of
the devices to determine the stage of operation thereof. Priority
means allocates priorities between the devices according to a
predetermined classification. In response to the outputs of the
monitoring means and priority means, an allocation means estimates
the amount of power consumption of the highest priority devices
during the remainder of the demand billing period and allocates the
estimated power thereto. The estimate is based upon the power
consumption of the present stage and the next stages and the
expected time to be spent in each. Subtracting means subtracts the
amount of allocated power from the total amount of power available
during the remainder of the demand billing period to thereby
determine the amount of remaining available power. In response to
the outputs of said monitoring means and said priority means,
allocation means similarly estimates the maximum amount of power
consumption of the remaining ones of said devices under normal
operation during the remainder of said demand billing period,
compares the estimated maximum amount of power consumption to the
amount of remaining available power determined by the subtracting
means to establish a ratio therebetween, and allocates said
remaining power to said remaining devices in accordance with said
ratio.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 comprises a graphical illustration of power consumption
during a demand period;
FIG. 2 comprises a diagrammatic illustration of the functions of a
power consumption control system for electric arc furnaces
employing the present invention;
FIG. 3 comprises an illustration of the kilowatt hour meter of FIG.
2;
FIG. 4 comprises a diagrammatic illustration of the relays of FIG.
2;
FIG. 5 comprises a diagrammatic illustration of a rheostat, control
mechanism and furnace of FIG. 2;
FIGS. 6 and 7 comprise an outline of the steps present in the
method of practicing the present invention; and
FIG. 8 comprises a diagrammatic illustration of computers 40 and 41
of FIG. 2.
DESCRIPTION OF THE PREFERRED EMBODIMENT
As stated above, electric utility companies normally employ a
demand billing technique for the billing of large customers. Hence,
under this system, it is to the customer's advantage to use the
same amount of power during all demand billing periods throughout
the working day.
FIG. 1 illustrates how kilowatt hour consumption of electricity
might vary during a demand period and still result in the same
total power consumption during the demand period. The straight line
10 represents a constant load throughout the demand period, curved
line 11 represents a load which is initially at a low value and
increases toward the end of the demand period, and curved line 12
represents a load which is initially at a high value and declines
toward the end of the demand period. A conventional power
consumption controller can operate well with the constant load 10
and the load represented by line 11, but would extrapolate the
initial load of line 12 in the manner shown by dashed line 13. The
conventional controller would thus assume that, without correction,
the limit would be reached during the initial part of the demand
period. The conventional controller would therefore reduce the
initial power consumption rate when the reduction was actually
unnecessary. The system of the present invention would, however,
predict the change in the process to occur later in the demand
period which will cause the total load during the demand period to
be less than the initial rate as shown by the subsequent portion of
line 12. The system of the present invention therefore would not
take control action to reduce consumption unless the limit would
still be exceeded.
Referring to FIG. 2, an exemplary power consumption control system
is shown employed with three electric arc furnaces 20-22 in a
foundry. It is noted that the situation of large varying and
controllable powerloads is typical of many industrial plants.
Therefore, this control system and method is not limited to
foundries and electric arc furnaces, but is equally applicable to
induction furnaces, annealing ovens, cathode plating operations and
many other types of industrial and electrical loads. Connected to
each furnace is a standard control mechanism 23-25. The control
mechanism comprises the means to control the furnace operation
under the direction of the melting foreman. Connected to each
control mechanism is a plurality of rheostats 26. The rheostats may
be controlled by the subject control circuitry or may be controlled
manually by the melting foreman. These rheostats cause the control
mechanism to adjust the spacing of carbon electrodes in the furnace
with respect to the surface of the steel, as will be explained
hereinafter. Steel in the furnace is melted by striking an arc from
the end of a carbon electrode to the surface of the steel. Raising
an electrode decreases current and power, and lowering an electrode
increases current and power. Each furnace is powered by a large
three-phase transformer, the power to which is supplied as a
portion of the power from a three-phase power supply input 27 from
the electric utility company. This power is supplied to a kilowatt
hour meter 28 and then on line 29 to the control mechanism for each
furnace.
Steel is made in the furnaces in batches which are called "heats."
A typical 10,000 -pound heat takes about 90 minutes to completely
process through a furnace, comprising six different stages of
operation. The stages, activity, typical time and power consumption
for each stage, including a seventh stage of furnace shutdown, are
shown below.
Typical Stage Activity Time Power 7 Furnace Shutdown Off 6 Tapping,
Patching, Charging 15 min. Off 5 Melt Down, A-Tap 35 min. On 4 Melt
Down, B-Tap 25 min. On 3 Oxygen Blow 3 min. Off 2 Heating, C-Tap 10
min. On 1 Alloys Added, C-Tap 2 min. On
The "A-Tap" refers to the 230 -volt tap from the power transformer
for a particular furnace, "B-Tap" refers to the 180 -volt tap from
the transformer, and "C-Tap" refers to the 110 -volt tap
therefrom.
Briefly, tapping the previous heat is the signal to start a new
heat. A heat is begun by the charging of the furnace with the
material to be melted, and is shown as stage 6. Upon charging,
turning the power on changes the status to stage 5 and an arc is
formed between the electrodes and the steel with the power supplied
from the A-Tap of the power transformer, thereby melting the steel.
A subsequent switch to the B-Tap of the power transformer changes
the heat to stage 4 as the melting continues. Then the power is
turned off and an oxygen blow occurs, comprising stage 3.
Restarting the arc from the C-Tap then comprises stage 2 and the
subsequent addition of alloys comprises stage 1. At this time, the
furnace returns to stage 6 and the heat is tapped.
The control mechanism 23-25 for each furnace includes a kilowatt
hour meter for supplying signals on cables 30-32. These signals are
applied to the monitor 35 of the subject power consumption control
system 36. The monitor includes circuitry for accumulating the
signals during a heat for each furnace and comparing the
accumulated number to preset numbers to thereby determine the stage
of operation of the furnace. The accumulation is reset to zero upon
the corresponding furnace beginning a new heat.
The priorities of the furnace are determined by the melting foreman
by means of a rotary switch which he manually sets to establish the
priority allocation in circuitry 37. The switch thus indicates
which of the three furnaces 20-22 has the highest priority. The
priority is determined by the melting foreman on the basis that the
next furnace to be tapped is given the highest priority. The
priority allocation might also be done automatically by merely
detecting that furnace which is in the lowest numbered stage. The
output of the priority allocation circuitry appears on lines 38 and
39 to high-priority power usage computer 40 and low-priority power
usage computer 41. This output comprises signals indicating the one
of the three furnaces which has been allocated the high
priority.
The operation of the circuitry comprising power control system 36
is accomplished under the control of timer 42. The timer is
responsive to utility synchronizing pulses supplied from input 43
on line 44. The synchronizing pulses are supplied from the utility
company at predetermined intervals, such as 15 minutes, to indicate
thereby the termination of one demand billing period and the
beginning of the next. The timer 42 comprises an interval timer
which transmits a signal to initiate operation of the power
consumption control circuitry upon receipt of the utility
synchronizing pulse and also subsequently supplies the signal 14
times at 1 -minute intervals. The signal from timer 42 is supplied
to monitor 35 and available power determination circuitry 45, to
initiate operation of the power consumption control circuitry 36.
The signal from timer 42 causes the comparison circuitry of monitor
35 to operate and supply the stage of operation of each furnace to
computers 40 and 41.
Available power determination circuitry 45 contains a counter which
is reset to zero upon the occurrence of a utility synchronizing
pulse from input 43 transmitted thereto on line 46. The counter is
incremented by pulses on line 47 from kilowatt hour meter 28. Each
pulse so supplied represents a specific increment of power usage
such as 22 kw. hr. Thus, the count present in the counter
represents the total amount of power utilized to that time during
the demand billing period. Circuitry 45 also includes a register
which is preset to the total amount of power which is to be
utilized during the demand billing period. The designation of this
amount is supplied at input 48 and is normally constant. The last
circuit comprising the available power determination circuitry 45
is a subtractor which operates upon occurrence of the signal from
timer 42 to subtract the count in the counter representing the
power usage up to that time from the contents of the register which
indicates the total amount of power to be used during the demand
period. The output of the subtractor represents the total power
available to the furnaces for the remainder of the demand period
and is supplied on output cable 49 to another subtraction circuit
50.
At the same time, priority allocation circuitry 37 supplies the
signals representing the highest priority furnace to computers 40
and 41 over lines 38 and 39. Computer 40 responds thereto by gating
data from monitor 35 giving the stage of operation of the
high-priority furnace. This data is used to check against data in
registers in computer 40 which also indicates the stage of
operation and the time remaining in that stage until the stage is
estimated to be completed. The stage number is translated by
internal circuitry to a power consumption rate and a time for
operation at that stage. Computer 40 includes a subtractor for
subtracting the time in the present stage for the furnace from the
time for the stage to thereby indicate the time remaining in the
present stage. This time is then compared by a logic circuit to the
time remaining in the demand period, which is obtained from the
timer 42. If the comparison shows that the time remaining in the
present stage is equal to or greater than the time remaining in the
demand period, this indicates that the furnace will remain in the
same stage during the entire remaining time of the demand period.
However, if the comparison shows that the time remaining in the
present stage is less than the time remaining in the demand period,
it indicates that the furnace will change from one stage to the
next during the remainder of the demand billing period.
The computer 40 also includes means for multiplying, which may
comprise either an electronic multiplier or an adder and means for
repetitive adding so as to accomplish the same result as
multiplication.
If the output of the logic circuit indicates that the furnace will
remain in the same stage for the remainder of the demand period,
the multiplying means operates to multiply the time remaining in
the demand period by the power consumption rate of the present
stage and supply the product at output T on cable 51.
However, if the output of the logic circuit indicates that the
furnace will go on to the next stage during the demand period,
computing circuitry 40 operates another multiplier to multiply the
time remaining in the present stage by the power consumption rate
of the present stage and to store that first partial product in a
register included in circuitry 40. Next, the circuitry operates a
subtraction circuit to subtract the time remaining in the preset
stage from the time remaining in the demand billing period. These
are the same amounts that were compared by the logic circuit,
above.
Referring to the table of stages, above, it is possible for the
furnace to operate in three stages during a single demand billing
period. Therefore, another comparison is made by another logic
circuit in the computer 40. In this instance, the amount of time
required for the next stage is compared by that logic circuit to
the time remaining in the demand billing period after completion of
the present stage, which is the result of the above
subtraction.
If the comparison shows that the time for the next stage is equal
to or greater than the subtracted time left in the demand period,
this indicates that the furnace will operate only in the present
and next stage during the remainder of the demand billing period.
Therefore, another multiplier in circuitry 40 is employed to
multiply the subtracted time left in the demand period by the power
consumption rate of the next stage. This product is then supplied
to an adder and added to the stored first partial product and the
total supplied at output T on cable 51.
However, if the output of the logic circuit shows that the time for
the next stage is less than the subtracted time left in the demand
period, this indicates that the furnace will operate in the
present, next and a third stage during the remainder of the demand
billing period. Therefore, another multiplier of circuitry 40 is
employed to multiply the time of the next stage by the power
consumption rate of that stage to form a second partial product
which is stored in a register included in circuitry 40. Next, the
time of the next stage is supplied to another subtraction circuit
of computer 40 and subtracted from the subtracted time left in the
demand period to form a remainder. The remainder is supplied to
another multiplier and multiplied by the power consumption rate for
the third stage. This product is then supplied to a three-way adder
and added to the first partial product and the sum added to the
second partial product. The resultant sum is then supplied at
output T on cable 51.
High-priority power usage estimation circuitry 40 also operates to
respond to the priority allocation signal on line 38 by supplying a
signal on one of the lines 52-54 to a bank of relays 55. The output
terminals of circuitry 40 to which lines 52-54 are connected are
labeled respectively A, B and C. The line at which the signal
appears indicates that that furnace is the highest priority
furnace. The relays of the bank of relays 55 operated by that
signal respond, as will be explained hereinafter, by supplying
signals on the set of lines 56-58 to thereby operate the rheostats
associated with the highest priority furnace. The operated
rheostats thus, in turn, cause the associated control mechanism
23-25 to operate the associated furnace 20-22 at maximum required
power.
Low priority power usage estimation circuitry 41 is identical to
the circuitry 40 except that it is arranged to respond to the
output of the priority allocation circuitry 37 on line 39 by
estimating the power consumption for the furnaces other than
high-priority furnace, and except that it includes an additional
register for storing the calculation for one furnace while
estimating the power usage for the other furnace. The estimated
power consumption for the two furnaces are then added and the
resultant total supplied at output terminal T on cable 60. At the
same time, circuitry 41 responds to the priority indication on line
39 by supplying a signal on each of the three lines 61-63 to power
ratio control 64. Output lines 61-63 are connected to output
terminals labeled A, B and C of circuitry 41 and indicate the two
furnaces that are not of the highest priority.
As mentioned above, the number indicating the amount of available
power for the furnaces for the remainder of the demand period is
supplied on cable 49 to subtractor 50. Also as discussed above, the
estimated power consumption for the high-priority furnace is
supplied on cable 51 to the subtractor 50. The subtractor operates
to subtract the number appearing on cable 51 from the number
appearing on cable 49 and supply the net result on cable 65. The
effect of this subtraction is to subtract the estimated
high-priority usage from the available power for the furnaces for
the remainder of the demand billing period. The net result
appearing on cable 65 is the remaining power available for the
low-priority furnaces.
As additionally discussed above, the number supplied on cable 60
represents the estimated power usage of the low-priority furnaces
if operated at maximum power. This output and the output of
subtractor 50 are supplied to division circuit 66. The division
circuit operates to divide the number on cable 65 by the number on
cable 60 and provide the resultant quotient on cables 67 and 68 to
low-priority power usage estimation circuitry 41 and to power ratio
control circuitry 64. This output quotient represents the ratio of
the available power for the low-priority furnaces to the power
usage by those furnaces if they were allowed to operate at 100
percent of power.
The signals on line 68 to power ratio control circuitry 64 operate
logic circuitry therein in accordance with the ratio designated by
the output signals to supply a signal on one of 11 lines. The 11
lines designate ratios at 10 percent intervals from 0 percent to
100 percent. These lines are each connected to three AND circuits,
each representing a furnace, of a matrix of 33 circuits. The other
input to each of the AND circuits is one of the lines 61-63. Thus,
only two of the 33 AND circuits are operated by a conjunction of
inputs, representing the ratio of power to be supplied to the two
low-priority furnaces. The output of each AND circuit is connected
to one of 33 output lines. These lines are grouped in cables
designated 70-72. The signals are thus provided on the two lines of
the same percentage in the cables 70-72 representing the two
low-priority furnaces. The signals are supplied to relays 55 which
transmit corresponding signals on cables 56-58 to rheostats 26. The
rheostats thus operate to insert a resistance in the line to the
control mechanism 23-25 for the corresponding furnace, which
operates as a potentiometer or voltage divider indicating the
percentage of power to be supplied to the corresponding furnace
20-22. This proportion will be the maximum required amount for the
highest priority furnace and a percentage of the required amount
for the low-priority furnaces which is designated by the output of
division circuitry 66.
Still referring to FIG. 2, after the operation of the control
mechanism 23-25, low-priority power usage estimation circuitry 41
employs the output of divider 66 on lines 67 by directly adding
that ratio to the previous time in stage for each low-priority
furnace and stores the total in the register representing time in
stage for that low-priority furnace. This ratio thus represents the
portion of a minute to be credited to the furnace for its operation
at the low power operation thereof. The circuitry 40 similarly adds
1 minute to the time in stage for the high-priority furnace. It is
of note that the time in stage stored in the circuitry 40 or 41 is
only an estimate and that the control mechanism 23-25 conducts the
actual operation of the furnace to change from one stage to
another. Thus, the output of the monitor 35 overrides the estimates
in circuitry 40 or 41.
The kilowatt hour meter 28 of FIG. 2 is shown in detail in FIG. 3.
The kilowatt hour meter is primarily a standard device as supplied
by the utility company. It includes a disk 80 mounted on a shaft
81. The shaft and disk rotate in accordance with the amount of
power being transmitted through the meter. The disk 80 includes a
plurality of slots 82. Associated with the disk is a light source
83 and a photodetector 84. The light emitted from light source 83
is shielded from the photocell 84 and impinges thereon only as a
slot 82 moves directly between the light source and photodetector.
The photodetector thus produces a pulse of electrical energy each
time a slot 82 is rotated between it and the light source 83. The
pulses produced by the photodetector are supplied to an amplifier
85. The amplifier amplifies the signal and supplies it to reed
relay 86. As each electrical pulse is supplied from the amplifier
85, the switch contacts 87 of the reed relay 86 are closed. A
filter comprising a resistor 88 and capacitor 89 serve to filter
out any contact bounce and allow the contact closing to be detected
on lines 47 of FIG. 2 by the counter of circuitry 45. The entire
meter with the exception of the filter 88, 89 is supplied by the
meter manufacturer. The filter is added only to prevent the counter
in circuitry 45 of FIG. 2 from counting the contact bounces and
thereby allows it to count only the generated pulses.
FIG. 4 illustrates one set of relays from the grouping of relays 55
in FIG. 2. Specifically, the lines 52-54 and the lines comprising
cables 70-72 are connected to specific ones of 11 sets of these
relays for each furnace. The lines 52-54 are each connected to the
corresponding 100 percent proportion relay, as is the 100 percent
line of each of the cables 70-72 to form an OR function. The
remaining relays are connected to the corresponding wires of cables
70-72. The connection is made at input terminal 90. The input is
connected to a reed relay 91 and it in turn is connected to a
positive power supply 92. Operation of the corresponding line thus
causes the signal to be transmitted to the reed relay 91, thus
closing the corresponding contacts 93. Closing these contacts
connects an AC power supply at terminals 94 and 95 to intermediate
relay 96. The intermediate relay is operated thereby to close the
corresponding contacts 97. Similarly, closing the contacts 97
connects the power supply at terminals 98 and 99 to power relay
100. Operation of the power relay causes closing of the
corresponding contacts 101 and thereby connects another power
supply at terminals 102 and 103 to the rheostat 26 of FIG. 2.
FIG. 5 illustrates the rheostats 26 for one of the furnaces, the
power circuit for the electrode one phase of that furnace and the
control mechanism for that furnace. For the purpose of
illustration, furnace 20 and control mechanism 23 are
illustrated.
The rheostat arrangement for each furnace comprises two linear
rheostats, a manual rheostat comprising resistance 108 and wiper
109 and an automatic rheostat comprising equal resistances 110-119
and relay contacts 120-130 of relays 131-141. Each of these relays
is connected to a power relay 100 of FIG. 4 as contained in the set
of relays 55 of FIG. 2.
Switching between the manual rheostat and the automatic rheostat is
accomplished by a switch which operates contacts 142-147. In the
manual position contacts 142-144 are closed and contacts 145-147
are open. In the automatic position contacts 145-147 are closed and
contacts 142-144 are open.
In the manual position, wiper 109 may be manually positioned so as
to operate the typical control mechanism 23. The voltage supply to
the rheostat is supplied between terminals 150 and 149 by the
control mechanism. The net voltage appearing between lines 148 and
149 is the inverse of the resistance and controls the operation of
the control mechanism. This much of the rheostat arrangement is
conventional. Other, corresponding arrangements may be employed to
operate other control mechanisms.
In the automatic position, the voltage presented to the control
mechanism 23 is determined by the one of the sets of contacts
120-130 which is closed by the associated relay 131-141. The one of
the sets of contacts which is closed causes the rheostat to present
the portion of the voltage from terminals 149 and 150 represented
by the relay at lines 148 and 149. This proportional voltage at
lines 148 and 149 thereby operates the control mechanism 23.
The typical control mechanism 23 includes a current-sensing means
151 arranged to sense the current in the powerline 152 supplying
each of the carbon electrodes 153-155 of the electric furnace 20
from powerlines 29. For the sake of illustration, only one of the
three powerlines for the three carbon electrodes will be shown.
The power line 152 is connected to a power switch 157 which is
controlled by control mechanism 23 to alternatively switch the
powerline between transformer taps 158, 159 and 160. As previously
discussed, these taps are called, respectively, A-Tap, B-Tap and
C-Tap, A-Tap having the highest voltage and C-Tap the lowest. Based
upon whether the powerline is connected to taps 158, 159 or 160,
the control mechanism 23 responds to the current sensed by
current-sensing means 151, the amount of which is supplied on line
156, to thereby transmit control signals on line 160 to hydraulic
mechanism 161. The signals on line 160 cause the hydraulic
mechanism to thereby raise or lower the corresponding electrode 153
to change the distance between the tip of the electrode and the
steel and thus change the current and power of the arc struck by
the electrode. The electrode is moved until the current sensed by
the current-sensing means 151 indicates to the control mechanism 23
that the proper power is being utilized by the electrode in
accordance with the setting of the manual rheostat 108, 109 or the
automatic rheostat 110-141.
The embodiment of the invention shown in FIG. 2 operates quite
effectively to control the power consumption of the three furnaces
20-22 to obtain constant power utilization for each power demand.
However, that embodiment comprising control circuitry 36 employs a
plurality of adding circuits, subtracting circuits and multiplying
circuits. From a cost of the machine standpoint, it would be much
more efficient to employ a single adder/subtractor and a single
multiple/driver which would be shared by the various functions
accomplished by the embodiment of FIG. 2. Such a machine would, of
course, comprise a general purpose digital computer. One example of
such a computer is the IBM 1800 machine. For the purpose of
allowing one having ordinary programming skills to program a
general purpose digital computer such as the IBM 1800 to accomplish
the method of the present invention, a flow diagram describing the
method of the present invention is illustrated in FIGS. 6 and 7.
The specific method of accomplishing the implementation of the
present invention is defined thereby with additional reference to
the following publications:
IBM 1800 Time-Sharing Executive Systems Specification Manual, Form
No. C26-590-0, 1964 ;
IBM 1800 Fortran Language Manual, Form No. C 26-5905-3, 1964 ;
and
IBM 1800 Assembly Language Manual, Form No. C26-5882-1, 1964 .
These publications are available from IBM Data Processing Division,
180 East Post Road, White Plains, New York 10601.
Referring to FIG. 6, the operation of the program is initiated by
an interval timer or utility synchronizing pulse, discussed with
respect to FIG 2. This is illustrated as block 200 in FIG. 6. The
IBM 1800, discussed above, includes an interval timer which may be
synchronized by the utility synchronizing pulses from input 43 in
FIG. 2. The interval timer, as discussed in the above manuals, can
then be set to initiate the program at 1 -minute intervals.
The program then proceeds to block 201 which instructs the IBM 1800
to read the kilowatt hour count for the overall plant operation. As
discussed in the above manuals, certain inputs of the IBM 1800 may
be connected to kilowatt hour meters to continuously count the
pulse train output of the kilowatt hour meters. The machine may
then be instructed to read the kilowatt hour count at specified
points in the program. The accumulated count is automatically reset
to zero by a program interrupt upon the occurrence of a utility
synchronizing pulse.
The next step in the program is represented by block 202 and
comprises an instruction to subtract the total amount of power
consumption this far in the demand billing period as obtained from
the kilowatt hour count in step 201 from a preset limit of power
consumption for the demand billing period. In the next step 203,
the computer is instructed to subtract the average uncontrolled
load, which is an estimate of the amount of power to be utilized
during the remainder of the demand period by the uncontrolled load,
from the result of step 202. The uncontrolled load concerned here
comprises the load on the power supply from the utility company
which comprises all other operations than the power employed by the
furnaces 20-22. This power includes such things as power to run the
computer, lights and air conditioning for the building, operation
of other machinery, etc. The uncontrolled power consumption
generally varies only slightly throughout the day and, hence, the
average thereof may be employed as an accurate estimate of future
usage. If this amount of power does not vary from day to day, a
preset standard factor may be employed by the computer for
subtraction. Otherwise, the computer may make averaging
calculations at various times.
Shown next is step 204 which comprises monitoring the status of
each furnace. Specifically, the IBM 1800 maintains the count of the
pulse train outputs from the kilowatt hour meters of the furnaces
as discussed above. The accumulated count for a furnace is
automatically reset to zero by a program interrupt as the furnace
begins a new heat. Step 204 comprises instructions to read the
accumulated kilowatt hour count for each of the furnaces, compare
each count to preset numbers to determine the stage of operation of
each furnace and store the indication thereof.
In step 205, the computer allocates priorities to the furnaces
based upon a manual input from the melting foreman. This input may
comprise the position of a manual switch supplied to one of the
inputs of the IBM 1800 computer.
Based upon the priority allocation, the computer proceeds to step
206 to estimate the power consumption for the remainder of the
demand billing period for the high-priority furnace. Block 206
comprises a plurality of steps which are illustrated in more detail
in FIG. 7.
The entry from step 205 to step 206 is shown by the "enter" block
207 of FIG. 7. In step 208, the computer compares the updated
storage stage of operation for the instant furnace, as will be
explained hereinafter, with the stage of operation therefore
detected by step 204 of FIG. 6. If the stages are not the same, the
output of step 204 is substituted for that obtained by the updating
process.
Proceeding to step 209, the computer then subtracts the updated
time in stage from the estimated total time for that stage, the
estimate having been kept in storage. The output from step 209
comprises the estimated remaining time in the present stage of
operation of the furnace.
In step 210, the estimated time remaining in the present stage is
compared to the time remaining in the demand period. If the result
of this comparison shows the time remaining in the present stage to
be equal to or greater than the time remaining in the demand
period, it indicates that the present stage of operation of the
furnace will occupy the entire remaining time of the demand period.
Therefore, the computer proceeds to step 211. If, however, the
result of the comparison shows it to be less, this indicates that
the furnace will enter a new stage during the remainder of the
demand billing period. In this case, the computer transfers to step
212.
In step 211, the remaining time in the demand billing period is
multiplied by the stored power consumption rate of the present
stage of the furnace. The product represents the estimate of the
power consumption of the high-priority furnace for the remainder of
the demand billing period. In step 213, this estimate is supplied
as the output for step 206 of FIG. 6.
Step 212 comprises multiplication of the remaining time in the
present stage by the power consumption rate of the present stage.
This product comprises the estimated power consumption of the
high-priority furnace for the completion of the present stage,
which amount is called the "1st partial power" and is temporarily
stored. The computer proceeds to step 214 wherein the inputs to
step 210 are again employed. The time remaining in the present
stage determined in step 209 is subtracted from the time remaining
in the demand billing period. The result of the subtraction is
called "X time" and represents the time remaining in the demand
billing period after completion of the present stage.
In step 215, the time for the next stage is compared to the X time.
If the result of this comparison shows the time for the next stage
to be equal to or greater than the X time, it indicates that the
next stage of operation of the furnace will occupy the entire final
part of the demand period. Therefore, the computer proceeds to step
216. If, however, the result of the comparison shows it to be less,
this indicates that the furnace will enter a third stage during the
final part of the demand period. In this case, the computer
transfers to step 217.
In step 216, the time remaining in the demand billing period after
completion of the present stage, this number comprising the X time
which is the result of the subtraction in step 214, is multiplied
by the power consumption rate of the next stage of the
high-priority furnace. This product and the 1st partial power
product stored from step 212 are then added in step 218. This total
therefore comprises the total estimated power consumption of the
high-priority furnace during the remainder of the demand billing
period. This total is then supplied as the output estimate in step
213 for the estimate of the high-priority power usage in step 206
of FIG. 6.
Step 217 comprises multiplication of the total time for the next
stage by the power consumption rate of that stage. This product
comprises the estimated power consumption for that stage of the
furnace and is called the "2nd partial power." The computer then
proceeds to step 219 and again employs the inputs to step 215,
subtracting the time of the next stage from X time. The result of
this subtraction is called "Y time" and represents the time
remaining in the final part of the demand billing period after
completion of the next stage. The next step 220 comprises the
multiplication of Y time by the power consumption rate of the third
stage of the demand period. This product and the 1st and 2nd
partial powers are then added in step 221. This total represents
the total estimated power consumption of the furnace during the
entire remainder of the demand billing period. This total is then
supplied as the output in step 213 for the estimate of the
high-priority power usage in step 206 of FIG. 6.
As is discussed in the above manuals, the IBM 1800 computer has the
capability of conducting and accomplishing several steps nearly
simultaneously. Therefore, at the same time that the computer is
conducting step 206 in FIG. 6, it is similarly conducting step 222.
Step 222 comprises the estimate of the power consumption of the
low-priority furnaces. The steps and procedures for making these
estimates are identical to those of step 206 and are the same as
illustrated in FIG. 7 and previously discussed.
At this time, the computer sets the rheostats 26 of FIG. 2 for the
high-priority furnace. The IBM 1800 computer is equipped with
digital outputs which are capable of supplying sufficient power to
operate the relays to control the associated control mechanism
23-25. The setting of the rheostats for the high-priority furnace
is shown as step 223 in FIG. 6. The computer then proceeds to step
224 and subtracts the output estimate from step 213 in FIG. 7 of
the power consumption for the high-priority furnace from the
available power as obtained from step 203 in FIG. 6. The result of
this subtraction is the estimated amount of remaining power
available for the low-priority furnaces. Proceeding to step 225,
the remaining power obtained as the output of step 224 is divided
by the estimated usage of the low-priority furnaces from step 222.
The result of this step comprises the ratio of available power to
the maximum power which would be used by the low-priority furnaces.
In step 226, the computer operates the relays 26 of FIG. 2 to
thereby set the rheostats in accordance with that ratio as
discussed with respect to the embodiment of FIG. 2.
In step 227, the computer updates the time in stage for each
furnace. For the high-priority furnace, this means adding 1 minute
to the stored time in stage, and for the low-priority furnaces the
ratio obtained from step 225 is directly added to the time in stage
stored therefor.
If at this time the time in stage becomes equal to that of the
stored total time for that stage, the computer moves to the next
stage of furnace operation and addresses the estimated total time
for that next stage for utilization by step 208 in FIG. 7. The
computer then proceeds to step 228 and terminates the program. The
program will then be initiated again by the interval timer as shown
by step 200.
It is therefore apparent that the system and method of the present
invention may be implemented either directly in controlling
hardware as shown in FIG. 2 or may be implemented by programming a
general-purpose digital computer such as the IBM 1800. This system
and method therefore accomplishes the true prediction of power
consumption for the furnaces for a demand billing period, employing
the stage of operation of each furnace and including in the
estimate any changes in states that may occur during the demand
billing period. The system and method therefore avoids employing
merely extrapolations of prior power consumption and more
accurately assures that the precise amount of power allocated for
each demand billing period will be exactly consumed.
For the purpose of illustration, an example of the circuitry
comprising either high-priority power usage computer 40 or
low-priority power usage computer 41 is shown in FIG. 8.
Briefly, the signals representing the highest priority furnace are
supplied on line 250 to control 251. Line 250 comprises line 38 for
computer 40 or line 39 for computer 41. In computer 40, the control
decodes the designation of the highest priority furnace and
transmits a signal on the corresponding one of the three outputs
comprising cable 252 to selectors 253 and 254. The three outputs
also comprise outputs 52-54 of FIG. 2. In computer 41, the control
251 decodes the designation of the highest priority furnace and
transmits signals on the two of the three outputs not corresponding
thereto. Selectors 253 and 254 respond to the control signal by
gating the inputs A, B or C to compare circuit 255. Selector 253 is
connected to the outputs of monitor 35 for the three furnaces and
selector 254 is connected to registers which store the updated
stage and time for the three furnaces. Compare circuitry 255 tests
for equality of the two stage designations for the furnace selected
by the two selectors. If the test shows equality, the output of
selector 254 is gated to stage and time storage register 256. If,
however, the stage designation from selector 253 is greater than
that from selector 254, the stage output of selector 253 is gated
to register 256 and the time of zero transmitted thereto. If the
test shows the stage designation of selector 253 to be less than
that from selector 254, then the stage output of selector 253 is
gated to register 256 and no time designation transmitted so as to
allow the register 256 to retain the prior time designation. In
this manner, that status of the furnace is checked for accuracy
with the monitor and corrected if necessary.
The stage output of register 256 is supplied to converter 257 which
translates the stage designation by gating a preset output selected
thereby. The preset output comprises a designation of the power
consumption rate for that stage at output R and a designation of
the time required for that stage at output T. The time for the
stage is thus supplied to one input of subtractor 258 and the time
spent in that stage is supplied from register 256 to the other
input of the subtractor. The subtractor subtracts input B from
input A and the output designates the time remaining in the present
stage and is supplied to the A input of logic circuit 259. The time
remaining in the demand period is supplied from timer 42 of FIG. 2
and temporarily stored in register 260. This time is supplied to
the B input of logic circuit 259 and compared to the A input. If
the A input is less, the logic circuit supplies a signal at output
L. If the A input is equal to or greater than the B input, a signal
is supplied at output G.
The signal at output G operates multiplier 261 to multiply the
power consumption rate for the present stage from output R of
converter 257 by the time remaining in the demand period from
register 260. The resultant product is supplied to output terminal
262 which corresponds to output T from computers 40 or 41.
A signal from output L of logic circuit 259 operates multiplier 263
and subtractor 264. Multiplier 263 multiplies the power consumption
rate of the present stage from converter 257 by the time remaining
in the present stage from register 256. Subtractor 264 subtracts
the time remaining in the present stage as supplied by subtractor
258 from the time remaining in the demand period supplied by
register 260 and supplies the result to the B input of logic
circuit 265.
At the same time, the stage output of register 256 is supplied to
incrementing circuit 266. Circuit 266 adds "1" to the received
stage designation and supplies the resultant designation of the
next stage to converter 267 and to incrementing circuit 268.
Circuit 268 operates similarly to thereby designate the following
stage to converter 269.
Converter 267 supplies the time for the next stage to the A input
of logic circuit 265. The logic circuit operates identically to
circuit 259. Thus, if the A and B inputs are equal or the A input
is greater, a signal is supplied at output G to operate multiplier
270. The multiplier multiplies the power consumption rate for the
next stage from converter 267 by the time left in the demand period
after the present stage supplied from subtractor 264. The product
is supplied to adder 271 and added to the power consumption of the
present stage supplied from multiplier 263. The resultant total
power consumption is supplied to output 262.
If input A to compare circuit 265 was less than the B input, a
signal is supplied at output L thereby to operate multiplier 272,
subtractor 273 and multiplier 274. Multiplier 272 multiplies the
power consumption rate of the next stage by the time for that
stage, both supplied by converter 267. The resultant product is
supplied to the B input of adder 275. Subtractor 273 subtracts that
time of the next stage at input A from the time remaining in the
demand period after completion of the present stage as supplied by
subtractor 264. The resultant time left for the third stage is then
supplied to multiplier 274 for multiplication by the power
consumption rate for that stage as designated by converter 269. The
resultant power consumption is supplied to input A of adder 275.
Input C to the adder comprises the power consumption of the present
stage from multiplier 263. The three power consumption amounts are
added by circuit 275 and supplied to output 262.
While the invention has been particularly shown and described with
reference to preferred embodiments thereof, it will be understood
by those skilled in the art that various changes in form and
details may be made therein without departing from the spirit and
scope of the invention.
* * * * *